Study protocol for collecting synchronized multimodal EMG, forearm kinematics and finger joint-angles data in healthy...

Rejection rates of upper-limb prosthetic hands remain high, largely due to limitations in control performance and usability. Most commercially available prosthetic hands rely on a fixed and limited set of predefined control modes, which restrict the execution of complex daily activities and reduce their acceptance by users. Consequently, there is a growing need for advanced control strategies that enable more natural and intuitive prosthetic hand operations. Multimodal sensing approaches that combine surface electromyography with inertial measurement unit and gyroscope data, together with accurate finger joint angle measurements, have shown promise for proportional and continuous control techniques. However, the availability of well-structured, synchronized datasets to support the development, training, and evaluation of such methods remains limited. This study aims to create a standardized, synchronized multimodal dataset to enable the benchmarking and training of proportional, continuous-control models for next-generation prosthetic hands and broader human-robot interaction and rehabilitation applications. The study employed a synchronized multimodal data acquisition framework, integrating surface EMG with kinematic measurements, enabling a comprehensive analysis of neuromuscular and movement data. Finger joint angles were obtained using an optical motion capture system without reliance on data gloves, with markers placed directly on the skin, which may introduce soft tissue artifacts and inter-subject variability.